Koelis Trinity launches ProMap Smart, an AI-assisted software that contours prostate cancer with 98% accuracy in seconds.
What ProMap Smart Adds to Prostate Contouring
Koelis Trinity has introduced ProMap Smart, an AI-assisted tool that contours prostate cancer targets in seconds and is described as reaching 98% accuracy. In clinical imaging workflows, contouring means drawing the boundaries of the prostate and suspicious regions so clinicians can plan biopsies, focal therapy, or follow-up imaging. That step is usually manual: a radiologist or urologist scrolls through multiparametric MRI or ultrasound-fusion volumes and marks structures slice by slice. The work is precise but repetitive, and small differences between readers can change target volumes, needle trajectories, and how lesions are tracked over time.
An automated contouring assistant does not replace clinical judgment. It proposes a first draft of the anatomy and lesion outline so the human expert can accept, edit, or reject it. The practical value is time and consistency: when the model is reliable, the clinician spends less effort on routine boundary drawing and more on decisions that still require experience—whether a region is truly suspicious, how aggressive the sampling plan should be, and how findings map to prior studies.
ProMap Smart sits in that category: AI-assisted contouring for prostate cancer, delivered as software associated with the Koelis Trinity platform, with speed measured in seconds and a stated accuracy level of 98%. Those two claims set expectations for what teams should verify in their own environment before relying on the output in routine care.
Where Contour Accuracy Matters in Practice
Prostate contouring is not only a drawing exercise. The outline defines volumes used for fusion-guided biopsy, treatment planning, and longitudinal comparison. If the contour expands too far into surrounding tissue, targets may include non-prostate structures and procedure plans become less focused. If it contracts too tightly, clinically relevant peripheral zones or lesions may be under-sampled. Accuracy at the level claimed for ProMap Smart matters most at the boundaries clinicians already debate: apex and base, transition versus peripheral zone, and the interface with seminal vesicles or neurovascular bundles when those regions affect planning.
Speed matters for different reasons. Contouring every case by hand can bottleneck busy imaging and interventional schedules. A model that returns a usable draft in seconds can keep fusion workflows moving when cases stack up. The tradeoff is familiar in medical AI: faster drafts only help if review is still rigorous. Teams that rubber-stamp model output risk encoding systematic errors; teams that re-draw everything erase the time gain. The useful operating mode is structured review—accept clean contours, edit ambiguous slices, and document corrections so the service line can see where the model struggles.
- Confirm that “accuracy” is measured against your gold standard (expert consensus, multi-reader study, or validated reference contours), not only vendor marketing language.
- Check performance on the sequences and fusion paths you actually use, including lower-quality studies and post-treatment anatomy.
- Require a human sign-off step so edited contours remain the clinical record of truth.
- Track edit rates over time; rising correction burden is a signal to retrain protocols or narrow use cases.
How Clinical Teams Should Evaluate a Tool Like This
Adoption should start with a limited pilot rather than immediate full-volume use. Run ProMap Smart on a representative set of recent cases already contoured by your specialists. Compare model output to existing contours for boundary agreement, missed extensions, and over-contouring. Measure wall-clock time from image availability to approved contour with and without the assistant. Include readers with different experience levels; a tool that helps junior staff more than seniors can still be valuable if senior review remains in place.
Also test failure modes explicitly. Contouring models often degrade when anatomy is distorted by prior procedures, when lesions sit at the extreme apex or base, or when image quality is poor. Document those cases and decide whether the software should be disabled, used only as a weak prior, or always forced through senior review. Because ProMap Smart is AI-assisted software rather than a fully autonomous diagnostic system, governance should treat it as a decision-support layer: version-controlled configuration, clear responsibility for final contours, and a path to report systematic issues back into quality review.
Fitting Auto-Contouring into an Existing Workflow
Successful integration depends less on the headline accuracy figure and more on where the tool sits in the day. Ideal placement is immediately after the diagnostic images are available and before biopsy planning or multidisciplinary discussion. The model proposes contours; the responsible clinician reviews them in the same viewer used for fusion or reporting; approved contours flow into the rest of the Koelis Trinity workflow without a second export step. Extra exports and format conversions are where time savings disappear.
Training should cover both the interface and the failure patterns. Staff need to know how to edit a contour efficiently, when to discard the suggestion entirely, and how to phrase reports so referring clinicians understand that boundaries were AI-assisted and human-verified. Over months, services that keep simple metrics—median review time, fraction of cases needing major edits, and concordance with biopsy outcomes where available—will know whether ProMap Smart is reducing drudgery without lowering contour quality. Used that way, a seconds-scale, high-accuracy contouring assistant can be a practical productivity tool rather than another black box in the imaging chain.